Activation of Transient Receptor Potential Canonical 3 (TRPC3)-mediated Ca2+ Entry by A1 Adenosine Receptor in Cardiomyocytes Disturbs Atrioventricular Conduction*

Background: A1-subtype of the adenosine receptors (A1AR) is arrhythmogenic. Results: A1AR activation enhanced Ca2+ entry through TRPC3 channel. Conclusion: TRPC3 is involved in conduction disturbances induced by A1AR. Significance: TRPC3 represents a promising target to prevent conduction disturbances. Although the activation of the A1-subtype of the adenosine receptors (A1AR) is arrhythmogenic in the developing heart, little is known about the underlying downstream mechanisms. The aim of this study was to determine to what extent the transient receptor potential canonical (TRPC) channel 3, functioning as receptor-operated channel (ROC), contributes to the A1AR-induced conduction disturbances. Using embryonic atrial and ventricular myocytes obtained from 4-day-old chick embryos, we found that the specific activation of A1AR by CCPA induced sarcolemmal Ca2+ entry. However, A1AR stimulation did not induce Ca2+ release from the sarcoplasmic reticulum. Specific blockade of TRPC3 activity by Pyr3, by a dominant negative of TRPC3 construct, or inhibition of phospholipase Cs and PKCs strongly inhibited the A1AR-enhanced Ca2+ entry. Ca2+ entry through TRPC3 was activated by the 1,2-diacylglycerol (DAG) analog OAG via PKC-independent and -dependent mechanisms in atrial and ventricular myocytes, respectively. In parallel, inhibition of the atypical PKCζ by myristoylated PKCζ pseudosubstrate inhibitor significantly decreased the A1AR-enhanced Ca2+ entry in both types of myocytes. Additionally, electrocardiography showed that inhibition of TRPC3 channel suppressed transient A1AR-induced conduction disturbances in the embryonic heart. Our data showing that A1AR activation subtly mediates a proarrhythmic Ca2+ entry through TRPC3-encoded ROC by stimulating the phospholipase C/DAG/PKC cascade provide evidence for a novel pathway whereby Ca2+ entry and cardiac function are altered. Thus, the A1AR-TRPC3 axis may represent a potential therapeutic target.

ity, and is a crucial regulator of the developing cardiovascular system. ADO derives from intra-and extracellular ATP degradation and accumulates when oxygen is lacking (1,2). Its action can be exerted through four adenosine receptor (AR) subtypes, namely A 1 , A 2A , A 2B , and A 3 .
Considerable evidence points to an important role for the ARs in protection against various myocardial injuries induced by hypoxia or ischemia in animal models and human, the cardioprotective action occurring mainly through activation of the A 1 AR (1)(2)(3)(4)(5)(6). However, other studies concede that A 1 AR contributes to arrhythmias including bradycardia, atrial fibrillation, conduction disturbances, and negative inotropy in the developing and adult heart (7)(8)(9)(10)(11).
For many years, the A 1 AR was considered as a G i proteincoupled receptor leading to inhibition of adenylyl cyclase and cAMP reduction, regulating cardiac function. Recent studies show that A 1 AR is coupled to dual signaling and stimulates the phospholipase C (PLC)/PKC pathway in smooth muscle cells and neurons and is currently identified as a G i/o protein-coupled receptor (12)(13)(14). It is also well established that A 1 AR can mediate modulation of cytosolic Ca 2ϩ concentration through two mechanisms such as direct regulation of plasmalemmal Ca 2ϩ channels or PLC-mediated mobilization of intracellular Ca 2ϩ stores via inositol 1,4,5-trisphosphate receptor (IP 3 R) in neurons and smooth muscle cells (15)(16)(17)(18). However, the role played by A 1 AR in Ca 2ϩ signaling in cardiomyocytes remains to be explored.
In addition to the well characterized mode of Ca 2ϩ entry through voltage-dependent Ca 2ϩ channels (e.g. L-and T-type Ca 2ϩ channels), receptor-mediated Ca 2ϩ -permeable cation channels activated by PLC are recognized for their physiological role (19).
In particular, the G protein-coupled receptor-mediated activation of the G q -PLC results in hydrolysis of phosphatidylinositol 4,5-bisphosphate with generation of the second messengers 1,2-diacylglycerol (DAG) and IP 3 , leading to IP 3 -induced release of Ca 2ϩ from endoplasmic and sarcoplasmic reticulum (ER/SR). The combined action of DAG and released Ca 2ϩ activate conventional PKCs whereas novel PKCs require only DAG. This signaling cascade activates plasmalemmal Ca 2ϩpermeable cation channels which are referred to receptor-operated channels (ROCs).
The transient receptor potential canonical (TRPC) channels have been postulated as the pore-forming proteins through which receptor-operated Ca 2ϩ entry (ROCE) occurs (20,21). There are seven members of the mammalian TRPC family, designated TRPC1-TRPC7, which assemble as homo-or heterotetramers to form cation-permeable channels. The properties of the heterotetramers are distinct from those of homotetramers. Using knocked down or knocked out strategies, TRPC channels have been originally proposed as store-operated channels (SOCs) activated by Ca 2ϩ depletion of stores (22)(23)(24). This situation remains highly controversial because of the recent identification of STIM1 (stromal interacting molecule 1) as an ER/SR Ca 2ϩ sensor and the Orai proteins forming the pore of SOC (25)(26)(27). In general, TRPC1, the first mammalian TRPC reported, can form heteromeric channels with TRPC4 and/or TRPC5 designated as SOCs, whereas TRPC3, TRPC6, and TRPC7 proteins, which share 75% identity, form ROCs and show activation sensitivity to the membrane-delimited action of DAG (23, 28 -31).
All isoforms except TRPC2 have been found at mRNA and/or protein levels in mammalian and avian cardiac muscle cells (32)(33)(34)(35)(36)(37) making them candidates for the receptor-operated nonselective cation channel known to exist in this cell type. There is accumulating evidence that TRPC channels mediate many physiological and pathological processes including arrhythmias, hypertrophy, heart failure, and apoptosis via ROCE (38). Indeed, a variety of studies using in vitro assays and transgenic and knock-out mice have suggested that TRPC3/6 proteins may assemble to form DAG-activated cation channels, which mediate G␣ q -mediated Ca 2ϩ signaling pathway. These TRPC-dependent pathways play a central role in the development of cardiac hypertrophy or arrhythmias (38 -42). We recently demonstrated that dysfunction of TRPC channels leads to second-degree atrioventricular blocks and ventricular arrhythmias in the embryonic chick heart model (37). In this model, the A 1 AR is expressed, and its activation is transiently arrhythmogenic through NADPH oxidase/ERK-and PLC/ PKC-dependent mechanisms whereas specific activation of A 2A AR, A 2B AR, or A 3 AR had no effect (11). The present study was designed to characterize the Ca 2ϩ entry pathway associated with the activation of A 1 AR in embryonic cardiac cells. In particular, the molecular mechanisms by which the TRPC channels could play a role in the A 1 AR-induced conduction disturbances have been investigated. Our findings reveal for the first time a new mechanism of TRPC3 channel activation dependent on A 1 AR activation and playing a predominant role in arrhythmogenesis.

EXPERIMENTAL PROCEDURES
Antibodies and Agents-Rabbit polyclonal antibodies used against TRPC1, 3, 4, 5, and 6 were from Alomone Labs (Jerusalem, Israel). Goat polyclonal anti-TRPC7 was from Everest Biotech (Oxfordshire, UK). The monoclonal antibody against cardiac troponin I (cTnI) was from Abcam (Cambridge, UK). Secondary antibodies for Western blotting were horseradish peroxidase-conjugated donkey anti-rabbit IgG (GE Healthcare) and horseradish peroxidase-conjugated goat anti-mouse IgG (Bio-Rad Laboratories). The specific agonist of A 1 AR CCPA, the L-type calcium channel inhibitor nifedipine, the general TRPC channels inhibitor SKF-96365 (SKF), the specific inhibitor of TRPC3 Pyr3, the PLC inhibitor U73122 and its inactive analog U73343, the DAG analog OAG, the PKC activator PMA, the general PKC inhibitor Ro 31-8220, and the irreversible SERCA inhibitor thapsigargin were from Sigma-Aldrich. The other general PKC inhibitor chelerythrine chloride, the myristoylated PKC pseudosubstrate inhibitor (MPI-PKC), the CRAC channel inhibitor BTP2, and the Fura-2/AM dye were from Calbiochem. The ER-targeted cameleon probe (D1 ER ) genetically targeted to the SR was used to determine specifically [Ca 2ϩ ] SR .
Atrial and ventricular myocytes were seeded on 0.1% gelatincoated glass coverslips in 35-mm plastic wells and expanded in culture for 3 days in the growth medium. These cultures were nearly pure as all cells showed contractile activity observed with phase-contrast microscope and expressed cTnI.
Cell Transfections-Atrial and ventricular myocytes were transiently transfected with an N-terminal fragment of hTRPC3 (amino acids 1-302 of TRPC3) cloned into pEYFP-C1 creating a N-terminally YFP-tagged fusion protein, 1 day after the seeding on glass coverslips using FuGENE® HD (Promega) according to the manufacturer's instructions. pmaxGFP (Amaxa Biosystems) was used as control to verify whether transfection itself could affect the Ca 2ϩ response. 2 g of plasmid (dominant negative (DN) TRPC3) was mixed in 250 l of serum-free medium, and 5 l of FuGENE® HD transfection reagent was added. The mixture was incubated for 15 min at room temperature and added drop by drop to the serum-free medium (1.75 ml) contained in each cell culture dish. After 24 h (at 37°C, 5% CO 2 ), the serum-free medium was replaced by the growth medium. The cells were kept in culture for further 24 -48 h until expression of the YFP fusion protein (as a marker of successful transfection) was detectable in the cells. Transfection efficiency was typically approximately 5%. Cardiomyocytes transfected with the DN-TRPC3 were selected via the YFP tag and subjected to measurement of CCPA-mediated Ca 2ϩ response.
For the D1 ER experiments, cells were transiently transfected with Lipofectamine® 2000 reagent (Invitrogen) by adding 2 g of cDNA/coverslip encoding the D1 ER construct. Cells were imaged 48 h after transfection.
Immunostaining-The cultured cells were fixed and permeabilized in cold 100% methanol for 5 min and washed three times in 1ϫ TBS containing 20 mM Tris, 154 mM NaCl, 2 mM EGTA, 2 mM MgCl 2 , pH 7.5. A saturation step was executed with TBS plus 1% BSA for 10 min. Samples were incubated for 1 h with primary antibody (monoclonal antibody against cTnI) diluted 1:200 in TBS plus 1% BSA. After washing out in TBS, cells were incubated for 1 h in TBS plus 1% BSA with secondary antimouse immunoglobulins labeled with Alexa Fluor 594 (Invitrogen). The cells were mounted using the Vectashield mounting medium with DAPI (Vector Laboratories, Burlingame, CA). The immunolabeled samples were examined using a conventional fluorescence microscope (Leica Microsystems).
Western Blotting-The lysates from atrial or ventricular myocytes and the whole heart were denatured, and 30 g of protein was loaded per lane, separated on SDS-polyacrylamide gels, and transferred to nitrocellulose membranes. Membranes were blocked and probed overnight with antibodies against TRPC1, 3, 4, 5, 6, and 7 and cTnI. After washes, the membranes were incubated with the secondary anti-rabbit, anti-goat, or anti-mouse IgG. Immunoreactive bands were detected with enhanced chemiluminescent procedure using the ECL Western blotting Analysis System (Amersham Biosciences). See supplemental Experimental Procedures for details.
PCR Amplification-Messenger RNA isolation from the whole embryonic heart or cultured atrial and ventricular myocytes obtained from 4-day-old chick embryos was performed using RNeasy Plus Mini kit (Qiagen). RT-PCR was conducted using SuperScript III one-step RT-PCR with platinum Taq (Invitrogen) in a Biometra TRIO-thermoblock (Lab Extreme Inc., Kent City, MI). See supplemental Experimental Procedures for details.
Measurement of Cytosolic Ca 2ϩ Changes Using Fura-2 Fluorescence-After 72 h in culture, atrial and ventricular myocytes were rinsed with a physiological solution containing 135 mM NaCl, 5 mM KCl, 2 mM CaCl 2 , 1 mM MgCl 2 , 10 mM HEPES, 10 mM D-glucose, pH adjusted to 7.4 with NaOH, and then incubated for 30 min in darkness in the same solution supplemented with 2 M Fura-2/AM (Invitrogen) plus 500 nM pluronic acid. Loaded cells were washed twice with the physiological solution before fluorometry. The changes in cytosolic Ca 2ϩ concentration were measured with Fura-2. Ratiometric images of Ca 2ϩ signal were obtained using an inverted microscope (Axio Observer, Zeiss) equipped with a Lambda DG4 illumination system (Sutter Instrument Company, Novato, CA), which alternatively changed the excitation wavelength between 340 nm (340AF15; Omega Optical) and 380 nm (380AF15). Emission was collected through a 415DCLP dichroic mirror, and a 510WB40 filter (Omega Optical), by a cooled, 12-bit CCD cam-era (CoolSnap HQ; Ropper Scientific Trenton, NJ) coupled to the microscope (x40 oil immersion fluorescence objective). Image acquisition in selected cells and analysis were performed with the Metafluor 6.3 software (Universal Imaging, West Chester, PA). To study the A 1 AR-activated Ca 2ϩ entry, the spontaneously beating cardiomyocytes to be explored were first selected in presence of external 2 mM Ca 2ϩ . Then, the cells were stimulated with 50 M CCPA, a specific agonist of A 1 AR, for 5 min in a Ca 2ϩ -free solution containing 1 mM EGTA. Subsequently, 2 mM Ca 2ϩ was re-added to the medium, and the peak amplitude of the fluorometric signal corresponding to the response to re-introduction of Ca 2ϩ was determined.
Measurement of Sarcolemmal Cation Influx Using Mn 2ϩ Quenching of Fura-2 Fluorescence-The procedure for loading cells and the set-up are the same as above. Fura-2 was excited at the isosbestic wavelengh, 360 nm, and emission fluorescence was monitored at 510 nm. The divalent cation influx was evaluated by the quenching of Fura-2 fluorescence when Mn 2ϩ ions enter into the cells. This technique exclusively reflects the cation influx through Ca 2ϩ channels. To study the A 1 AR-activated Ca 2ϩ entry, the cardiac cells were stimulated with 50 M CCPA in Ca 2ϩ -free medium, and then 500 M Mn 2ϩ was re-added to the medium with Ca 2ϩ . The quenching rate of fluorescence intensity (F) was estimated using linear regression of the initial decaying phase (slope, ⌬F/⌬t) just after Mn 2ϩ addition and expressed as the decrease of F per min, normalized to the maximal F signal obtained before Mn 2ϩ (100%) to correct for differences in the cell size and/or fluorophore loading. Photobleaching was Ͻ0.5%/min during measurements.
Measurement of SR Ca 2ϩ Changes-Atrial and ventricular myocytes were transiently transfected with cDNA encoding the D1 ER construct 48 h before the experiments. Cardiomyocytes were illuminated at 440 nm (440AF21; Omega Optical), and emission was collected through a 455DRLP dichroic mirror, alternatively at 480 nm (480AF30; Omega Optical) and 535 nm (535AF26; Omega Optical). Photometric values were corrected for photobleaching and expressed as ratio 535/480. Values were normalized to the signal obtained before CCPA.
Ex Vivo Mounting of the Heart and Experimental Protocol-The isolated spontaneously beating hearts were placed in each well of 24-plastic wells containing 1 ml of the medium (Ϯdrugs) and stabilized for 45 min at 37.5°C on the thermostabilized stage of an inverted microscope (Leica DMI3000 B, Wetzlar, Germany) as described previously (37). The mean atrial beating rate of control and treated hearts was measured every 5 min for 60 min, and all arrhythmias were noted. See supplemental Experimental Procedures for details.
Electrocardiogram of the Whole Heart-Intact spontaneously beating hearts were placed in the culture compartment of a stainless steel air-tight chamber maintained at 37.5°C. ECG displayed characteristic P, QRS, and T components, which allowed us to determine the beating rate from PP or RR interval (beats/min), PR interval (ms), QT duration (ms), and QRS complex duration (ms) as described previously in details (37,43). See supplemental Experimental Procedures for details.
Statistical Analysis-All values are reported as mean Ϯ S.E. For all experiments, the significance of any difference between two groups was assessed with one-way analysis of variance (ANOVA) completed by Tukey's post hoc test. The statistical significance was defined by a value of p Յ 0.05 (*, p Յ 0.05; **, p Յ 0.01; ***, p Յ 0.001).

Characterization of Atrial and Ventricular Myocytes-After
72 h of culture, cardiomyocytes isolated from atria or ventricle of 4-day-old chick embryos differentiated into spontaneously beating atrial and ventricular myocytes at a rhythm of 106 Ϯ 6 (n ϭ 20) and 57 Ϯ 9 (n ϭ 18) beats/min at 37°C, respectively. All myocytes displayed the characteristic spindle-shaped morphology (Fig. 1A), with sarcomeric organization. The majority of myocytes was positive for the cTnI immunostaining (Fig. 1B), and Western blotting showed the presence of cTnI in both cultured cells and whole heart (Fig. 1C). Only atrial and ventricular myocytes with spontaneous beats were investigated.
TRPC Channels and A 1 AR Are Expressed in Atrial and Ventricular Myocytes-In cultured atrial and ventricular myocytes, Western blotting analysis identified protein expression of TRPC1, 3-7 (TRPC2 being expressed only in rodents (44)), and RT-PCR revealed transcript for A 1 AR (Fig. 2). The whole heart was used as a positive control for TRPC proteins as previously published (37). For each TRPC isoform, a fusion protein was used as control antigen for negative control (data not shown). Western blotting for A 1 AR was not performed because no antibody matching the chicken is available on the market.
A 1 AR Activation Mobilizes Ca 2ϩ Mainly via TRPC3 Isoform-Under control conditions, atrial and ventricular myocytes showed a basal Ca 2ϩ entry (Fig. 3, A and B). Stimulation of A 1 AR by a specific agonist CCPA (50 M) (45) did not induce SR Ca 2ϩ depletion in Ca 2ϩ -free medium but doubled Ca 2ϩ entry in the presence of extracellular Ca 2ϩ with respect to basal level in both types of cardiomyocytes. Blockade of TRPC channels by a widely used inhibitor at 40 M SKF (46) strongly reduced the CCPA-induced Ca 2ϩ entry, suggesting that A 1 ARinduced Ca 2ϩ entry is mediated by TRPCs.
To confirm the A 1 AR-induced Ca 2ϩ response and for the rest of the study, we have used an alternative approach that measured sarcolemmal divalent cation entry using Mn 2ϩ quenching of Fura-2 fluorescence method into Fura-2-loaded cells. This technique takes advantage of the fact that essentially all Ca 2ϩ -permeable cation channels exhibit permeability to Mn 2ϩ . The rate of fluorescence quenching after the addition of Mn 2ϩ showed that, in the absence of A 1 AR stimulation, there was a basal cation entry inhibitable by SKF in atrial and ventricular myocytes (supplemental Fig. S1). Inhi-   3 primary cultures). B, A 1 AR mRNA was identified by RT-PCR in cultured myocytes and whole heart (n ϭ 2 primary cultures). PCR product of the predicted size was 318 bp. ␤-Actin was amplified as a positive control. The negative control (Ctrl(Ϫ)) contained water instead of DNA. AUGUST   vated Ca 2ϩ channel, and the L-type Ca 2ϩ channel, participate in the basal cation influx.

TRPC3 Channel Activation by A 1 Adenosine Receptor
In agreement with the data shown in Fig. 3, A and B, the CCPA-enhanced cation entry in atrial and ventricular myocytes was inhibited by SKF (Fig. 3, C and D). Furthermore, BTP2 and Pyr3 abolished the CCPA-induced cation entry, confirming that TRPC channels are involved in adenosinergic signaling and, in particular, the TRPC3 isoform in atrial myocytes (Fig.  4A). Surprisingly, in ventricular myocytes, although Pyr3 also strongly reduced the cation influx activated by CCPA, BTP2 had no effect. To further support the contribution of TRPC3 in A 1 AR-enhanced cation entry, we transiently transfected cardiomyocytes to express an N-terminal fragment of TRPC3, which exerts a DN effect on TRPC3 channel function presumably due to disruption of channel assembly (48 -50). A suffi-cient number of atrial and ventricular myocytes were efficiently transfected as shown with the co-staining for the tag-YFP and cTnI to perform the experiment (supplemental Fig. S2). Expression of YFP-DN-TRPC3 significantly suppressed CCPA-induced cation entry in atrial and ventricular myocytes (Fig. 4A), supporting the concept that TRPC3 constitutes a key element in adenosinergic signaling. The empty vector (pmaxGFP) used as control had no effect on A 1 AR-induced cation influx. It should be noticed that the cation influx is the result of two components: one insensitive to Pyr3 (the basal cation entry in unstimulated cells as shown in supplemental Fig. S1), which represented approximately 40% of the total CCPA-induced cation influx, and the other sensitive to Pyr3 or to overexpression of the DN-TRPC3 (Fig. 4A). These findings show that approximately 85% of A 1 AR-dependent cation influx is carried by

TRPC3 channel in atrial and ventricular myocytes, once basal influx is deduced.
Additionally, we showed by cytosolic Ca 2ϩ measurement that A 1 AR activation did not induce Ca 2ϩ release from internal stores. To corroborate this result, we transfected the myocytes with a genetically encoded Ca 2ϩ probe targeted to the SR (D1 ER ) to detect Ca 2ϩ store depletion after A 1 AR activation. Accordingly, the SR Ca 2ϩ depletion was almost undetectable with the D1 ER after A 1 AR stimulation by CCPA in Ca 2ϩ -free medium, whereas the irreversible SERCA inhibitor thapsigargin strongly depleted the Ca 2ϩ store in atrial and ventricular cells (Fig. 4B). These results suggest that adenosine in-duced Ca 2ϩ entry through TRPC3 in a store-independent mechanism.
A 1 AR-dependent Cation Entry through TRPC3 Channel Involves PLC/DAG/PKC Pathway-We first examined whether A 1 AR activation stimulates PLCs which are known to activate TRPC channels. Inhibition of PLCs by U73122, indeed, significantly reduced the CCPA-induced cation entry in cardiomyocytes. The inactive analog of this inhibitor, U73343, did not affect the cation influx (Fig. 5A). Activation of PLCs is known to induce the cleavage of phosphatidylinositol 4,5-bisphosphate into DAG and IP 3 . In turn, DAG activates the conventional and novel PKC isoforms. Based on their requirements for activa- In A-C, cation entry was normalized to that induced by CCPA alone. In A and B, all inhibitors, activators, U73343, and OAG were added 3 min before CCPA. In C, the myocytes were pretreated with the peptide MPI-PKC 30 min before CCPA. tion, three PKC classes are defined: conventional cPKCs (␣, ␤, ␥), which are activated by Ca 2ϩ and DAG; novel nPKCs (␦, ⑀, , ), which require DAG but are Ca 2ϩ -independent; and atypical aPKCs (, , ), which are activated independently of Ca 2ϩ or DAG. The fact that two inhibitors of all PKC isoforms, chelerythrine and Ro 31-8220 (Ro 31) significantly reduced the CCPA-induced cation entry showed that PKCs are also involved in A 1 AR-mediated Ca 2ϩ influx in atrial and ventricular myocytes (Fig. 5A).
When PLCs were inhibited by U73122, preventing the formation of DAG and, in turn, activation of conventional and novel PKCs, the DAG analog (OAG) restored 53% of the cation entry whereas the activator of conventional and novel PKCs (PMA) had no effect in atrial myocytes (Fig. 5B), suggesting that DAG can directly activate TRPC channels independently of PKCs. By contrast, in ventricular myocytes, either OAG or PMA restored A 1 AR-induced cation entry by 50% or 69%, respectively, suggesting that DAG acts indirectly on TRPC channels via PKCs. Furthermore, SKF (supplemental Fig. S3) and Pyr3 (Fig. 5B) abolished the OAG-and/or PMA-induced restoration of cation entry in both type of cells, indicating that TRPC3 channel activity is mostly regulated by DAG and/or PKCs.
Additionally, the direct DAG-dependent regulation of TRPC3, the absence of SR Ca 2ϩ depletion, and the absence of effect of PMA on A 1 AR-induced cation entry in atrial myocytes suggest that an atypical PKC isoform could be involved in the adenosine-dependent Ca 2ϩ response. The MPI-PKC inhibited the A 1 AR-mediated Ca 2ϩ entry by 86% in atrial myocytes and only by 56% in ventricular myocytes (Fig. 5C). The ventricular myocytes were less sensitive to MPI-PKC than atrial myocytes, suggesting a predominant contribution of PKC in A 1 AR-mediated Ca 2ϩ entry in atrial cells.
TRPC3 Isoform Is Involved in A 1 AR-induced Conduction Disturbances-The proportion of spontaneously arrhythmic hearts was always Ͻ20% under basal (ctrl) condition (Fig. 6, A  and B). ADO and CCPA induced transient arrhythmias after 5 min in 65 and 80% of the hearts, respectively. SKF at 5 M as well as Pyr3 at 10 M attenuated ADO-and CCPA-induced arrhythmias (Fig. 6, A and B). These findings indicate that overactivation of TRPC channels, in particular of the TRPC3 isoform contribute to A 1 AR-induced arrhythmias.
We also explored the contribution of TRPC channels in CCPA-induced arrhythmias on the basis of electrocardiography (ECG) of the whole heart. The effects of CCPA, SKF, and Pyr3 alone or combined on atrial and ventricular beating rate, atrioventricular conduction (PR interval), ventricular activation (QT duration), and intraventricular conduction (QRS complex width) were determined ex vivo. It should be noticed that the electrical parameters were stable for at least 60 min under control conditions (data not shown). The important changes in ECG morphology and alteration of functional parameters induced by CCPA, CCPAϩSKF, and CCPAϩPyr3 are illustrated in Figs. 6 and 7. CCPA rapidly induced seconddegree atrioventricular blocks, essentially in the form of Mobitz type I (Wenckebach phenomenon) characterized by a progressive lengthening of the PR interval followed by a dropped QRS without alteration of the atrial rhythm (Figs. 6Cb and 7b). It should be noticed that the period of arrhythmias induced by CCPA varied from one heart to another ranging from 5 to 45 min as we observed previously (11).
The administration of 5 M SKF in the presence of CCPA rapidly suppressed the conduction disturbances induced by A 1 AR stimulation for at least 60 min (Fig. 6C, c and d). As expected, a slight transient bradycardic effect of CCPA was observed after 5 min (Fig. 6D) without affecting PR interval and QRS duration (data not shown) for at least 60 min as we showed previously (11). CCPAϩSKF or SKF alone increased the QT duration from 30 min onward (Fig. 6, E and F), which was due to the specific effect of SKF alone on ventricular activation as recently observed (37).
The same pattern of response was obtained in presence of Pyr3 which rapidly suppressed the conduction disturbances in all hearts (Fig. 7, c and d), suggesting that TRPC3 has a major role in A 1 AR-induced conduction abnormalities. The noticeable protection against these arrhythmias afforded by blockade of all TRPC channels by SKF or by specific inhibition of TRPC3 isoform by Pyr3 is illustrated by ECG tracings shown in supplemental Fig. S4.

DISCUSSION
In the present work, we assessed the role of TRPC3-encoded ROC in the A 1 AR-induced cardiac conduction disturbances. We showed that activation of A 1 AR by CCPA increased sarcolemmal Ca 2ϩ entry without SR Ca 2ϩ depletion in embryonic atrial and ventricular myocytes. This A 1 AR-induced Ca 2ϩ entry was mainly due to upstream activation of PLCs and PKCs because it was prevented by U73122 and chelerythrine or Ro 31-8220, respectively. Although the isoforms IP 3 R1, 2, and 3 were expressed at mRNA level in atria and ventricle of the embryonic heart (except IP 3 R2 in atria; data not shown), there was no Ca 2ϩ release from SR measured with Fura-2 fluorescence in Ca 2ϩ -free solution and no detectable change of SR intraluminal Ca 2ϩ measured with the D1 ER probe after exposure to CCPA, indicating that SR Ca 2ϩ release was not associated with A 1 AR stimulation. In adult cardiac fibers, the contribution of A 1 AR to adenosine-induced Ca 2ϩ release from SR is also negligible compared with A 2A R (51). Because A 1 AR is coupled to PLC at the (sub)sarcolemmal level, it is possible that IP 3 production was not sufficient to optimally activate the sarcoplasmic IP 3 Rs and lead to a measurable Ca 2ϩ release and/or that IP 3 could not reach IP 3 Rs as suggested in arterial myocytes (52). It is also conceivable that, at the investigated stage of development, the IP 3 Rs could not be functional. Therefore, the effects of A 1 AR activation being strictly dependent on extracellular Ca 2ϩ could be attributed to Ca 2ϩ entry through sarcolemmal ROC by a store-independent mechanism.
Additionally, in atrial and ventricular myocytes pretreated with the inhibitor of PLCs, the DAG analog OAG evoked Ca 2ϩ entry indicating the involvement of a DAG-dependent Ca 2ϩ influx. Interestingly, in atrial myocytes, when PLCs were inhibited, OAG restored the Ca 2ϩ entry induced by CCPA whereas the activator of conventional and novel PKCs (PMA) had no effect. By contrast, in ventricular myocytes, either OAG or PMA restored Ca 2ϩ entry. These results indicate that, in atrial cells, A 1 AR stimulation induced Ca 2ϩ entry via a G protein coupled to PLC and that DAG formation plays a pivotal role in TRPC3 Channel Activation by A 1 Adenosine Receptor AUGUST 3, 2012 • VOLUME 287 • NUMBER 32 this phenomenon via a PKC-independent mechanism as suggested in CHO and smooth muscle cells (28,53,54). This is in contrast to the Ca 2ϩ entry in ventricular cells which was stimulated by DAG via PKCs, as reported in portal vein myocytes (55).
As PKCs have been postulated to play a key role in adenosine signaling (11,56,57), subsequent focus was placed on the PKC pathway. In our model, general PKC inhibitors (Ro 31-8220 and chelerythrine) indeed abolished the effect of A 1 AR stimulation on Ca 2ϩ entry in both atrial and ventricular myocytes. In ventricular myocytes, the novel PKC isotypes appeared to be principally involved in adenosinergic signaling as DAG but not Ca 2ϩ is required for activation of this PKC family. In adult cardiomyocytes, activation of A 1 AR promotes targeting of the novel PKC isoforms to caveolin-rich plasma membrane microdomains resulting in their activation (57). These observations are consistent with our findings that A 1 AR activated novel PKCs which are known as the most abundant Ca 2ϩ -independent PKC isoforms in neonatal and adult ventricular cardiomyocytes (58). Surprisingly, in atrial myocytes, PKCs were activated neither by OAG and PMA nor by Ca 2ϩ , suggesting that atypical PKC isoforms are involved in adenosinergic response. PKC is the most abundant isoform in fetal myocardium, but the mode of activation and its role in cardiac function are not completely understood (59). The selective inhibitor of PKC (MPI-PKC) strongly decreased the A 1 AR-induced plasmalemmal Ca 2ϩ influx in atrial cells (Ϫ86%) whereas this inhibitor had a slighter effect (p Ͻ 0.01) in ventricular cells (Ϫ56%), supporting a predominant role for PKC in adenosinergic signaling in atrial cells. Adenosine A 2A receptor activation has been shown to induce translocation/activation of PKC (60,61), and the G q protein can be regarded as a scaffold protein capable of recruiting PKC into a complex that activates the ERK pathway in neonatal and adult cardiomyocytes and fibroblasts (62). We also recently found that A 1 AR activation induces pacemaking and conduction disturbances through downstream activation of ERK pathway in the developing heart (11). Thus, these observations and our present data indicate that, in parallel to the DAG-dependent Ca 2ϩ influx, A 1 AR activation may be able to recruit PKC which, in turn, could regulate ROC in atrial myocytes and, to a lesser extent, in ventricular cells.
The closely related candidates for ROC are TRPC3/6/7 channels subfamily which can be activated in response to DAG in a membrane-delimited action independently of Ca 2ϩ store depletion (28,63,64). Because Ca 2ϩ mobilization in embryonic atrial and ventricular cells was store-independent, we hypothesized that ROC could be constituted of TRPC proteins, especially TRPC3 subfamily because of its activation by DAG.
All TRPC isoforms at the protein level, except TRPC2, were expressed in cultured atrial and ventricular myocytes. The fact that SKF significantly reduced the A 1 AR-, PMA-, and/or OAGinduced Ca 2ϩ influx clearly indicates that TRPC channels play a role in the adenosinergic response. The other drug BTP2 used in this study is known to inhibit TRPC3, 5, and 6 (40, 65) and Orai channels (CRAC and SOC), which are activated by Ca 2ϩ store depletion through a STIM-dependent mechanism (66,67). In our model, BTP2 reduced A 1 AR-mediated Ca 2ϩ influx in atrial but not in ventricular myocytes, which could be attributable to distinct types of assembly of TRPC isoforms into homo-or heterotetramers ion channels and/or to their differential interactions with STIM/Orai proteins.
TRPC3 isoform is known to be strongly implicated in various pathophysiological processes. Indeed, TRPC3 protein forms DAG-activated Ca 2ϩ channel, which mediates G q -in-  AUGUST 3, 2012 • VOLUME 287 • NUMBER 32 duced hypertrophy in rat neonatal cardiomyocytes and transgenic mice overexpressing TRPC3 (68 -70). In adult ventricular cardiomyocytes submitted to ischemia, activation of the purinergic P2Y 2 receptor by ATP/UTP activates heteromeric TRPC3/7 channels leading to cell depolarization, Ca 2ϩ overload, and arrhythmias (39). The fact that A 1 AR-enhanced Ca 2ϩ entry was completely abolished by selective inhibition of TRPC3 by Pyr3 and significantly reduced by the DN TRPC3 construct preventing TRPC3 channel assembly indicates that TRPC3 predominantly contributed to the ROC pathway in atrial and ventricular cells. Furthermore, Pyr3 significantly reduced the PMA-and/or OAG-mediated Ca 2ϩ influx, clearly indicating that TRPC3 isoform was regulated by DAG and PKCs in atrial and ventricular myocytes. Our hypothesis that TRPC3 isoform is the predominant component of the A 1 AR-stimulated Ca 2ϩ entry is also strengthened by the fact that TRPC3 is generally activated through activation of PLC and DAG production.

TRPC3 Channel Activation by A 1 Adenosine Receptor
Thus, A 1 AR, via PLCs activation, triggers a nonselective cationic influx occurring through TRPC3 channel. However, we cannot rule out the possibility that A 1 AR activation can stimulate other TRPC isoforms because TRPC3 can form functional ROC channels also as heterotetramers.
We showed previously that (i) adenosine induces transient arrhythmias through A 1 AR whereas specific activation of A 2A AR, A 2B AR, or A 3 AR had no effect and (ii) activation of A 1 AR is pro-arrhythmic through NADPH oxidase/ERK-and PLC/PKC-dependent mechanisms (11). These findings are in accordance with our present data showing the involvement of PLCs, DAG, and PKCs in the TRPC3-dependent Ca 2ϩ entry triggered by A 1 AR stimulation in atrial and ventricular myocytes. Moreover, it has been reported that Ca 2ϩ entry through TRPC3 channel regulates ERK phosphorylation via PKC activation (71), a mechanism that could occur also in our model via PKC isoform. Electrocardiography of the whole heart showed that A 1 AR activation induced rapidly and transiently second degree atrioventricular blocks (essentially in the form of Wenckebach phenomenon) and that selective inhibition of TRPC3 channel by Pyr3 immediately suppressed these arrhythmias. It should be noticed that the well known transient bradycardia induced by an adenosinergic stimulation after 5 min was exclusively due to A 1 AR activation in our model and not to TRPC channels. By contrast, the SKF-induced prolongation of the QT duration observed after 30 min was due to inhibition of TRPCs (suppressing the negative regulation of Cav1.2 channel by TRPCs in ventricle as we previously documented (37) by a mechanism independent of adenosine signaling).
As in adult heart, a tight control of intracellular Ca 2ϩ level is required to maintain normal cardiac activity in the developing heart because any activation of transmembrane Ca 2ϩ influx can result in Ca 2ϩ overload associated with arrhythmias and contractile dysfunction (72,73). Thus, on the basis of our findings, we propose that a subtle activation/inhibition of voltage-independent Ca 2ϩ channels like TRPC3 could play a crucial role in Ca 2ϩ homeostasis and consequently in regulation of electromechanical activity of the developing heart. Any alteration of these channels by activation of G protein-coupled receptors (such as ARs) under pathological conditions could affect Ca 2ϩ entry and promote pro-arrhythmic alterations of membrane potential. For instance, activation of TRPC3/6 channels by G␣ q and DAG results in early afterdepolarizations in the failing heart of adult mouse (42).
We demonstrated previously that TRPC channels play a key role in regulating cardiac pacemaking, conduction, and ventricular activity without any stimulation of GPCR. Here, we highlight a novel axis involving specifically A 1 AR/TRPC3 in conduction abnormalities. The present study provides the first demonstration that A 1 AR activation in cardiomyocytes elicits ROCE which is mediated by the TRPC3 channel isoform via PLC-, DAG-, and PKC-dependent mechanisms (Fig. 8). Furthermore, an enhanced TRPC3-dependent Ca 2ϩ entry can lead to a rise of intracellular Ca 2ϩ concentration and trigger transient arrhythmias in the developing heart model. Such an A 1 AR-dependent signal transduction could play a crucial role FIGURE 8. Proposed model of A 1 AR-induced ROCE through TRPC3 in atrial (A) and ventricular (B) myocytes. The A 1 AR-induced Ca 2ϩ entry through TRPC3 channel requires the upstream activation of PLC/DAG pathway in atrial and ventricular myocytes. The atypical PKC predominant in atrial myocytes and the novel PKCs predominant in ventricular myocytes are crucial for regulating TRPC3 activity. Increased ROCE via TRPC3 appears to be involved in conduction disturbances induced by A 1 AR stimulation in the developing heart. PIP2, phosphatidylinositol 4,5-bisphosphate; aPKC, atypical PKC; nPKC, novel PKC.
in rhythm and conduction disturbances observed under hypoxia or ischemia when ADO accumulates in the interstitial fluid of the myocardium. Hence, the TRPC3 channel isoform may be regarded as a new potential therapeutic target to reduce intracellular Ca 2ϩ overload and subsequent arrhythmias in fetal and adult heart.